[Wiki Loves Monuments] Minor API improvement

[André Costa]

My plan wasn't for an Android hack but rather a change to how the
wlm-database matches the entries of the Wikipedia row templates (for
se-fornmin). I'm assuming the mapping is done on a case by case basis
anyway since the different datasets will include different fields all with
different parameter names. In this way the change should fix the Swedish
entries (independently of platform) without changing anything for anyone
else.

Based on Commons:Monuments_database/Statistics the only other database with
few names is de-he which "only" has names for ~45% of the entries (by
comparison se-fornmin has 6.25%). The reason I wanted the parameter
combination is that the entries are commonly referred to by the
{{{raä-nr}}} but only some have a common name (the {{{namn}}} parameter).
In the other Swedish database, se-bbr, these were both entered into the
{{{namn}}} parameter of the row template to start with, I only split them
for se-fornmin because last year we had issues with standardisation in how
this should be presented.

Adding the type parameter ({{{typ}}}) to the database name would help
clarify what it is that the person is actually looking for, e.g. a
runestone, a burial mound, a ruin, i.e. the sort of thing that is normally
obvious from the common name where there is one.

Cheers,
Andre

On 7 September 2012 22:11, Platonides <platonides at gmail.com> wrote:

> On 07/09/12 19:29, Arthur Richards wrote:
> >>From a quick search through the monument database, there appear to be
> > ~165k monuments with no name:
> (...)
> > All but one of those monuments has data in the 'field':
> > mysql> select count(*) from monuments_all where name='' and source='';
> > +----------+
> > | count(*) |
> > +----------+
> > |        1 |
> > +----------+
>
> Source is the wikipedia page. Of course all monuments have one, but that
> isn't too useful.
>
> That odd monument is in country India, in English, adm1 IN-UP. No other
> data.
>
>
>
> > If it's too difficult to otherwise determine a name for the monument,
> > perhaps the data could be updated to construct a name based off of the
> > 'title' field in the source, perhaps with an incremented digit appended
> > to it. For instance:
> >
> http://de.wikipedia.org/w/index.php?title=Liste_der_denkmalgeschützten_Objekte_in_Sedliská&redirect=no&useskin=monobook&oldid=105722097
> > could become:
> > Liste der denkmalgeschützten Objekte in Sedliská - 01
> >
> > Or something like that. Rather than hack something into the android app
> > for this, I recommend that this happen in the database itself. That way
> > everyone using the app, including older versions of the app, would be
> > able to see the updated monuments.
>
> Rather than the full page title, you'd probably want something like
> "Sedliská - Unnamed 01" or so. But it seems something better done in the
> local lists than blindly at the db.
>
>
> The problem is that sometimes the official lists from the government
> show a monument but not the name. So you know that there's some monument
> in that city, the id it should have, but absolutely no idea of what it
> is (of course coordinates are even less likely to appear).
> And it'd be very hard to find it out. I think that those monuments
> should simply not be displayed by the app (by default, at least).
>
> They are of interest for the wikipedia lists, but quite useless for WLM.
> I don't think anyone could come with a photo for them.
>
> If they do have an id (all but that odd Indian entry do), another option
> would be to use "Unknown monument with id <id>" as name.
>
> _______________________________________________
> Wiki Loves Monuments mailing list
> WikiLovesMonuments at lists.wikimedia.org
> https://lists.wikimedia.org/mailman/listinfo/wikilovesmonuments
> http://www.wikilovesmonuments.org
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.wikimedia.org/mailman/private/wikilovesmonuments/attachments/20120907/cd5f670d/attachment.html>